Project Summary Overdose deaths attributable to prescription opioid analgesics (opioids) in the United States have more than quadrupled since 1999, and are now the leading cause of death among those under 50 years of age in the United States. Patients receiving legitimate opioid prescriptions account for a large proportion of those who misuse opioids, transition to injecting drugs, and are at high risk for overdose and death as well as blood-borne infections. Therefore, tapering patients off of long-term opioid prescriptions is now viewed as a high priority. However, tapering carries its own risks as inadequately controlled pain is a key predictor of opioid misuse and transition to heroin and other injection drug use. In response, the CDC recently published dosing guidelines for physicians of chronic non-cancer pain patients that identify avoidance, tapering, and discontinuation of prescription opioids when possible as high priorities. However, the guidelines related to tapering and discontinuation are essentially a clinical ?best guess? because of the absence of high-quality data. The long-term goal of our research program is to improve opioid prescribing strategies to minimize risks of substance misuse, overdose, and blood-borne infections. The overall objective of this application is to leverage a unique combination of ?big data? and qualitative data to generate high-quality evidence about optimal strategies to taper or end long-term opioids. We will achieve this objective by leveraging a powerful combination of insurance claims, health system, cause of death, and blood-borne infection diagnosis data covering ten years and ~1 million residents of North Carolina supported by complementary qualitative interviews. We hypothesize that chronic non-cancer pain patients whose high-dose opioid prescriptions are rapidly tapered (more quickly than suggested by CDC guidelines) or abruptly terminated will be at higher risk for overdose and blood-borne infection than patients whose doses are gradually tapered, controlling for key differences between these groups. We further hypothesize that engagement in mental health or substance misuse treatment (e.g. buprenorphine) when doses are being reduced will be protective against adverse health outcomes. This 2-year R21 application provides a unique opportunity to use large-scale data and robust methods to address the absence of data on the risks of tapering and termination strategies for long-term opioid prescriptions. In addition to informing evidence-based guidelines, results from this observational study can be directly incorporated into a subsequent R01 application to conduct a sequential multiple assignment randomization trial (SMART) to experimentally evaluate the risks and benefits of one or more tapering strategies identified in this project.